I don't think it's a shoddy design as much as shoddy documentation. I would be proud to create a generator that works as well as this one. But supplying no diagram of a major component is inexcusable. Forcing a sincere technician to reverse engineer an expensive unit isn't good documentation.
Having said that, I have many other major complaints about the documentation. On the surface it appears that they made a good effort but when you try to follow the flow of signals and commands it falls short. Locating major parts is another deficiency.
It's not just this unit; I am going nuts with a 3455A as well. My 3456A was obtained at a good price because the HPIB function was intermittent, causing the unit to stop working. Luckily, that has fixed itself and it seems to work reliably now. My 5328A had to be cobbled up from three units to make one good one, and I am happy. There is still one issue but it's not causing problems for me.
Since the latter unit only works up to about 600 MHz I bought a cheapie Chinese 2.4 GHz counter and am mostly satisfied with it. For just a few bucks it has given me good value and service. While it has a TCXO, it's no match for the oven oscillators in the generator and counter. But if I need a good calibration and can work reasonably quickly, I can compare it at a lower frequency and then measure, making appropriate correction.
The hobby has to include repair. But it's a learning experience.
I am eager to see the results of your testing.